|
40744
|
1506
|
0
|
2026-05-14T09:15:57.892478+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-14/1778 /Users/lukas/.screenpipe/data/data/2026-05-14/1778750157892_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: use concurrency = true
ggml_me ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:43:13.536522Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=3899909268894052485, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:43:16.247373Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:43:21.435499Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8439548885587338707, trigger=click)
2026-05-14T11:43:21.504514Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)
2026-05-14T11:43:29.518808Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)
2026-05-14T11:43:45.924284Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:16.951333Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:18.317865Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:18.386225Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:20.185352Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:20.265226Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:21.430482Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:21.480628Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:22.304274Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 117 eligible frames
2026-05-14T11:44:25.629527Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 60 frames, 3.7MB → 0.0MB (75.5x), 60 JPEGs deleted
2026-05-14T11:44:26.154304Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:26.200137Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:28.743329Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 55 frames, 8.8MB → 0.2MB (37.7x), 55 JPEGs deleted
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
2026-05-14T11:45:16.431532Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:45:16.483574Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:45:18.618752Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:45:19.474033Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:45:29.158075Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=593730046027495091, trigger=click)
2026-05-14T11:45:29.200661Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:45:32.974859Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:45:33.023926Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:45:42.120191Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:45:42.228078Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:46:38.666105Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:46:38.718980Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:47:10.843681Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:47:22.053549Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:48:24.882731Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:48:24.898344Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:48:43.049629Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 93.0ms elapsed (expected 5.3ms) → inserting 87.7ms silence (8420 samples)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:49:25.141660Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:49:33.255012Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=90 elapsed=4.502131875s
2026-05-14T11:49:33.255192Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 90 eligible frames
2026-05-14T11:49:35.892354Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 2.9MB → 0.0MB (61.0x), 47 JPEGs deleted
2026-05-14T11:49:38.390817Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 41 frames, 6.5MB → 0.2MB (28.3x), 41 JPEGs deleted
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:51:28.386976Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:52:16.643026Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:52:16.672181Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:53:30.916512Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:54:04.647564Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:54:04.697193Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:54:38.546214Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 45 eligible frames
2026-05-14T11:54:39.871380Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 1.4MB → 0.0MB (30.4x), 22 JPEGs deleted
2026-05-14T11:54:41.821322Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 21 frames, 3.4MB → 0.2MB (14.6x), 21 JPEGs deleted
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:55:36.908674Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:57:39.484346Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:58:59.457881Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T11:58:59.481640Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:09.423657Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:09.474124Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:11.661414Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:11.718517Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:59:42.205405Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:59:44.810433Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=80 elapsed=2.975872084s
2026-05-14T11:59:44.811786Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 80 eligible frames
2026-05-14T11:59:47.272581Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 43 frames, 2.6MB → 0.0MB (56.3x), 43 JPEGs deleted
2026-05-14T11:59:49.607250Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 35 frames, 5.6MB → 0.2MB (24.2x), 35 JPEGs deleted
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-14T12:00:24.589833Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:25.655040Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:26.832387Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:38.456296Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:40.235853Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:40.833256Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:42.087724Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:42.177013Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:01:46.599136Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:03:49.206275Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T12:04:54.719219Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=23 elapsed=5.038572166s
2026-05-14T12:04:54.720627Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 23 eligible frames
2026-05-14T12:04:55.702295Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 11 frames, 0.7MB → 0.0MB (15.6x), 11 JPEGs deleted
2026-05-14T12:04:56.674389Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 1.6MB → 0.2MB (7.0x), 10 JPEGs deleted
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:05:52.348533Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:07:54.829608Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T12:08:35.035877Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)
2026-05-14T12:08:35.051180Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)
2026-05-14T12:09:03.069084Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)
2026-05-14T12:09:03.070997Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:09:57.680290Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
2026-05-14T12:10:00.073643Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=22 elapsed=3.386172375s
2026-05-14T12:10:00.073940Z INFO screenpipe_engine::snapshot_compaction: snapshot compactio...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:43:13.536522Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=3899909268894052485, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:43:16.247373Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:43:21.435499Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8439548885587338707, trigger=click)\n2026-05-14T11:43:21.504514Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)\n2026-05-14T11:43:29.518808Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)\n2026-05-14T11:43:45.924284Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:16.951333Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:18.317865Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:18.386225Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:20.185352Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:20.265226Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:21.430482Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:21.480628Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:22.304274Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 117 eligible frames\n2026-05-14T11:44:25.629527Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 60 frames, 3.7MB → 0.0MB (75.5x), 60 JPEGs deleted\n2026-05-14T11:44:26.154304Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:26.200137Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:28.743329Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 55 frames, 8.8MB → 0.2MB (37.7x), 55 JPEGs deleted\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\n2026-05-14T11:45:16.431532Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:45:16.483574Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:45:18.618752Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:45:19.474033Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:45:29.158075Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=593730046027495091, trigger=click)\n2026-05-14T11:45:29.200661Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:45:32.974859Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:45:33.023926Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:45:42.120191Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:45:42.228078Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:46:38.666105Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:46:38.718980Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:47:10.843681Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:47:22.053549Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:48:24.882731Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:48:24.898344Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:48:43.049629Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 93.0ms elapsed (expected 5.3ms) → inserting 87.7ms silence (8420 samples)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:49:25.141660Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:49:33.255012Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=90 elapsed=4.502131875s\n2026-05-14T11:49:33.255192Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 90 eligible frames\n2026-05-14T11:49:35.892354Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 2.9MB → 0.0MB (61.0x), 47 JPEGs deleted\n2026-05-14T11:49:38.390817Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 41 frames, 6.5MB → 0.2MB (28.3x), 41 JPEGs deleted\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:51:28.386976Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:52:16.643026Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:52:16.672181Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:53:30.916512Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:54:04.647564Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:54:04.697193Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:54:38.546214Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 45 eligible frames\n2026-05-14T11:54:39.871380Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 1.4MB → 0.0MB (30.4x), 22 JPEGs deleted\n2026-05-14T11:54:41.821322Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 21 frames, 3.4MB → 0.2MB (14.6x), 21 JPEGs deleted\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:55:36.908674Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:57:39.484346Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:58:59.457881Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:58:59.481640Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:09.423657Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:09.474124Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:11.661414Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:11.718517Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:59:42.205405Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:59:44.810433Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=80 elapsed=2.975872084s\n2026-05-14T11:59:44.811786Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 80 eligible frames\n2026-05-14T11:59:47.272581Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 43 frames, 2.6MB → 0.0MB (56.3x), 43 JPEGs deleted\n2026-05-14T11:59:49.607250Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 35 frames, 5.6MB → 0.2MB (24.2x), 35 JPEGs deleted\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T12:00:24.589833Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:25.655040Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:26.832387Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:38.456296Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:40.235853Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:40.833256Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:42.087724Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:42.177013Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:01:46.599136Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:03:49.206275Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T12:04:54.719219Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=23 elapsed=5.038572166s\n2026-05-14T12:04:54.720627Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 23 eligible frames\n2026-05-14T12:04:55.702295Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 11 frames, 0.7MB → 0.0MB (15.6x), 11 JPEGs deleted\n2026-05-14T12:04:56.674389Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 1.6MB → 0.2MB (7.0x), 10 JPEGs deleted\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:05:52.348533Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:07:54.829608Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T12:08:35.035877Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:08:35.051180Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:09:03.069084Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:09:03.070997Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:09:57.680290Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\n2026-05-14T12:10:00.073643Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=22 elapsed=3.386172375s\n2026-05-14T12:10:00.073940Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 22 eligible frames\n2026-05-14T12:10:00.162432Z WARN sqlx::query: summary=\"PRAGMA wal_checkpoint(TRUNCATE)\" db.statement=\"\" rows_affected=1 rows_returned=1 elapsed=1.17871275s\n2026-05-14T12:10:01.105437Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 0.6MB → 0.0MB (14.3x), 10 JPEGs deleted\n2026-05-14T12:10:02.014368Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 1.6MB → 0.2MB (7.0x), 10 JPEGs deleted\n2026-05-14T12:11:39.174895Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:11:39.176743Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:11:54.721845Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 90.5ms elapsed (expected 5.3ms) → inserting 85.2ms silence (8178 samples)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:12:03.340132Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:14:07.520305Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\n2026-05-14T12:15:04.735575Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=1 rows_returned=113 elapsed=2.624938375s\n2026-05-14T12:15:13.762879Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=66 elapsed=11.72677475s\n2026-05-14T12:15:13.765115Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 66 eligible frames\n2026-05-14T12:15:15.812779Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 33 frames, 2.0MB → 0.0MB (44.3x), 33 JPEGs deleted\n2026-05-14T12:15:18.036326Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 31 frames, 4.9MB → 0.2MB (21.5x), 31 JPEGs deleted","depth":4,"on_screen":true,"value":"ggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:43:13.536522Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=3899909268894052485, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:43:16.247373Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:43:21.435499Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8439548885587338707, trigger=click)\n2026-05-14T11:43:21.504514Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)\n2026-05-14T11:43:29.518808Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)\n2026-05-14T11:43:45.924284Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:16.951333Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:18.317865Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:18.386225Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:20.185352Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:20.265226Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:21.430482Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:21.480628Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:22.304274Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 117 eligible frames\n2026-05-14T11:44:25.629527Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 60 frames, 3.7MB → 0.0MB (75.5x), 60 JPEGs deleted\n2026-05-14T11:44:26.154304Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:26.200137Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:44:28.743329Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 55 frames, 8.8MB → 0.2MB (37.7x), 55 JPEGs deleted\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\n2026-05-14T11:45:16.431532Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:45:16.483574Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:45:18.618752Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:45:19.474033Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:45:29.158075Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=593730046027495091, trigger=click)\n2026-05-14T11:45:29.200661Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)\n2026-05-14T11:45:32.974859Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:45:33.023926Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:45:42.120191Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:45:42.228078Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:46:38.666105Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:46:38.718980Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:47:10.843681Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:47:22.053549Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:48:24.882731Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:48:24.898344Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:48:43.049629Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 93.0ms elapsed (expected 5.3ms) → inserting 87.7ms silence (8420 samples)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:49:25.141660Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:49:33.255012Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=90 elapsed=4.502131875s\n2026-05-14T11:49:33.255192Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 90 eligible frames\n2026-05-14T11:49:35.892354Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 2.9MB → 0.0MB (61.0x), 47 JPEGs deleted\n2026-05-14T11:49:38.390817Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 41 frames, 6.5MB → 0.2MB (28.3x), 41 JPEGs deleted\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:51:28.386976Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:52:16.643026Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:52:16.672181Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:53:30.916512Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:54:04.647564Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:54:04.697193Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)\n2026-05-14T11:54:38.546214Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 45 eligible frames\n2026-05-14T11:54:39.871380Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 1.4MB → 0.0MB (30.4x), 22 JPEGs deleted\n2026-05-14T11:54:41.821322Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 21 frames, 3.4MB → 0.2MB (14.6x), 21 JPEGs deleted\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:55:36.908674Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:57:39.484346Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:58:59.457881Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:58:59.481640Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:09.423657Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:09.474124Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:11.661414Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T11:59:11.718517Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T11:59:42.205405Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T11:59:44.810433Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=80 elapsed=2.975872084s\n2026-05-14T11:59:44.811786Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 80 eligible frames\n2026-05-14T11:59:47.272581Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 43 frames, 2.6MB → 0.0MB (56.3x), 43 JPEGs deleted\n2026-05-14T11:59:49.607250Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 35 frames, 5.6MB → 0.2MB (24.2x), 35 JPEGs deleted\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T12:00:24.589833Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:25.655040Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:26.832387Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:38.456296Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:40.235853Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:40.833256Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:42.087724Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)\n2026-05-14T12:00:42.177013Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:01:46.599136Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:03:49.206275Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T12:04:54.719219Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=23 elapsed=5.038572166s\n2026-05-14T12:04:54.720627Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 23 eligible frames\n2026-05-14T12:04:55.702295Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 11 frames, 0.7MB → 0.0MB (15.6x), 11 JPEGs deleted\n2026-05-14T12:04:56.674389Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 1.6MB → 0.2MB (7.0x), 10 JPEGs deleted\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:05:52.348533Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:07:54.829608Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T12:08:35.035877Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:08:35.051180Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:09:03.069084Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:09:03.070997Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:09:57.680290Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\n2026-05-14T12:10:00.073643Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=22 elapsed=3.386172375s\n2026-05-14T12:10:00.073940Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 22 eligible frames\n2026-05-14T12:10:00.162432Z WARN sqlx::query: summary=\"PRAGMA wal_checkpoint(TRUNCATE)\" db.statement=\"\" rows_affected=1 rows_returned=1 elapsed=1.17871275s\n2026-05-14T12:10:01.105437Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 0.6MB → 0.0MB (14.3x), 10 JPEGs deleted\n2026-05-14T12:10:02.014368Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 1.6MB → 0.2MB (7.0x), 10 JPEGs deleted\n2026-05-14T12:11:39.174895Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:11:39.176743Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)\n2026-05-14T12:11:54.721845Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 90.5ms elapsed (expected 5.3ms) → inserting 85.2ms silence (8178 samples)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:12:03.340132Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T12:14:07.520305Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\n2026-05-14T12:15:04.735575Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=1 rows_returned=113 elapsed=2.624938375s\n2026-05-14T12:15:13.762879Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=66 elapsed=11.72677475s\n2026-05-14T12:15:13.765115Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 66 eligible frames\n2026-05-14T12:15:15.812779Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 33 frames, 2.0MB → 0.0MB (44.3x), 33 JPEGs deleted\n2026-05-14T12:15:18.036326Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 31 frames, 4.9MB → 0.2MB (21.5x), 31 JPEGs deleted","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.0674867,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.33776596,"top":1.0,"width":0.06732048,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33976063,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.40508643,"top":1.0,"width":0.06732048,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.40708113,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.47240692,"top":1.0,"width":0.06732048,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.4744016,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5397274,"top":1.0,"width":0.06732048,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.54172206,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.60704786,"top":1.0,"width":0.06732048,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6090425,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6743683,"top":1.0,"width":0.06732048,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.67636305,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-8787432842042246226
|
-1970252785695001119
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: use concurrency = true
ggml_me ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:43:13.536522Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=3899909268894052485, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:43:16.247373Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:43:21.435499Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8439548885587338707, trigger=click)
2026-05-14T11:43:21.504514Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)
2026-05-14T11:43:29.518808Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8439548885587338707, trigger=click)
2026-05-14T11:43:45.924284Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:16.951333Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:18.317865Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:18.386225Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:20.185352Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:20.265226Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:21.430482Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:21.480628Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:22.304274Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 117 eligible frames
2026-05-14T11:44:25.629527Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 60 frames, 3.7MB → 0.0MB (75.5x), 60 JPEGs deleted
2026-05-14T11:44:26.154304Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:26.200137Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:44:28.743329Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 55 frames, 8.8MB → 0.2MB (37.7x), 55 JPEGs deleted
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
2026-05-14T11:45:16.431532Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:45:16.483574Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:45:18.618752Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:45:19.474033Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:45:29.158075Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=593730046027495091, trigger=click)
2026-05-14T11:45:29.200661Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1242313268223121183, trigger=click)
2026-05-14T11:45:32.974859Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:45:33.023926Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:45:42.120191Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:45:42.228078Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:46:38.666105Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:46:38.718980Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:47:10.843681Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:47:22.053549Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:48:24.882731Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:48:24.898344Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:48:43.049629Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 93.0ms elapsed (expected 5.3ms) → inserting 87.7ms silence (8420 samples)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:49:25.141660Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:49:33.255012Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=90 elapsed=4.502131875s
2026-05-14T11:49:33.255192Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 90 eligible frames
2026-05-14T11:49:35.892354Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 2.9MB → 0.0MB (61.0x), 47 JPEGs deleted
2026-05-14T11:49:38.390817Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 41 frames, 6.5MB → 0.2MB (28.3x), 41 JPEGs deleted
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:51:28.386976Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:52:16.643026Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:52:16.672181Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:53:30.916512Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:54:04.647564Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=9006240032176046078, trigger=click)
2026-05-14T11:54:04.697193Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=9006240032176046078, trigger=click)
2026-05-14T11:54:38.546214Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 45 eligible frames
2026-05-14T11:54:39.871380Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 1.4MB → 0.0MB (30.4x), 22 JPEGs deleted
2026-05-14T11:54:41.821322Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 21 frames, 3.4MB → 0.2MB (14.6x), 21 JPEGs deleted
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:55:36.908674Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:57:39.484346Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:58:59.457881Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T11:58:59.481640Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:09.423657Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:09.474124Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:11.661414Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T11:59:11.718517Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T11:59:42.205405Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T11:59:44.810433Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=80 elapsed=2.975872084s
2026-05-14T11:59:44.811786Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 80 eligible frames
2026-05-14T11:59:47.272581Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 43 frames, 2.6MB → 0.0MB (56.3x), 43 JPEGs deleted
2026-05-14T11:59:49.607250Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 35 frames, 5.6MB → 0.2MB (24.2x), 35 JPEGs deleted
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-14T12:00:24.589833Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:25.655040Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:26.832387Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:38.456296Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:40.235853Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:40.833256Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:42.087724Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=5219237514711016516, trigger=click)
2026-05-14T12:00:42.177013Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=5219237514711016516, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:01:46.599136Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:03:49.206275Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T12:04:54.719219Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=23 elapsed=5.038572166s
2026-05-14T12:04:54.720627Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 23 eligible frames
2026-05-14T12:04:55.702295Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 11 frames, 0.7MB → 0.0MB (15.6x), 11 JPEGs deleted
2026-05-14T12:04:56.674389Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 10 frames, 1.6MB → 0.2MB (7.0x), 10 JPEGs deleted
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:05:52.348533Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:07:54.829608Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T12:08:35.035877Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)
2026-05-14T12:08:35.051180Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)
2026-05-14T12:09:03.069084Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=1801589742647015621, trigger=click)
2026-05-14T12:09:03.070997Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1801589742647015621, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T12:09:57.680290Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
2026-05-14T12:10:00.073643Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=22 elapsed=3.386172375s
2026-05-14T12:10:00.073940Z INFO screenpipe_engine::snapshot_compaction: snapshot compactio...
|
40742
|
NULL
|
NULL
|
NULL
|
|
45861
|
NULL
|
0
|
2026-05-14T15:31:17.975891+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-14/1778 /Users/lukas/.screenpipe/data/data/2026-05-14/1778772677975_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:07:25.268127Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=10 elapsed=1.819536167s
2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames
2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted
2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)
2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:22:46.614239Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=1 rows_returned=0 elapsed=4.94553125s
2026-05-14T18:22:52.962854Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=392 elapsed=1.0767815s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:07:25.268127Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=10 elapsed=1.819536167s\n2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames\n2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted\n2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:07:25.268127Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=10 elapsed=1.819536167s\n2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames\n2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted\n2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (nc)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-1176716547952649155
|
-816202924460018208
|
click
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:07:25.268127Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=10 elapsed=1.819536167s
2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames
2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted
2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)
2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:22:46.614239Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=1 rows_returned=0 elapsed=4.94553125s
2026-05-14T18:22:52.962854Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=392 elapsed=1.0767815s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv ...
|
45860
|
NULL
|
NULL
|
NULL
|
|
45863
|
1637
|
82
|
2026-05-14T15:31:18.860686+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-14/1778 /Users/lukas/.screenpipe/data/data/2026-05-14/1778772678860_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:07:25.268127Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=10 elapsed=1.819536167s
2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames
2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted
2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)
2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:22:46.614239Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=1 rows_returned=0 elapsed=4.94553125s
2026-05-14T18:22:52.962854Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=392 elapsed=1.0767815s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:07:25.268127Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=10 elapsed=1.819536167s\n2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames\n2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted\n2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:07:25.268127Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=10 elapsed=1.819536167s\n2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames\n2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted\n2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (nc)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-1176716547952649155
|
-816202924460018208
|
visual_change
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:07:17.813880Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:07:25.268127Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=10 elapsed=1.819536167s
2026-05-14T18:07:25.268371Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 10 eligible frames
2026-05-14T18:07:25.824320Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 5 frames, 0.6MB → 0.2MB (4.0x), 5 JPEGs deleted
2026-05-14T18:07:26.273059Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 3 frames, 0.7MB → 0.4MB (1.7x), 3 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:09:20.299383Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)
2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:22:46.614239Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=1 rows_returned=0 elapsed=4.94553125s
2026-05-14T18:22:52.962854Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=392 elapsed=1.0767815s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv ...
|
45862
|
NULL
|
NULL
|
NULL
|
|
45865
|
1640
|
0
|
2026-05-15T06:26:36.297147+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826396297_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)","depth":4,"on_screen":true,"value":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-8190670393490586440
|
-852231723626465824
|
manual
|
accessibility
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ ...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45866
|
1641
|
0
|
2026-05-15T06:26:36.353658+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826396353_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
3799306250804398193
|
-852231723626465824
|
manual
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45867
|
1640
|
1
|
2026-05-15T06:26:41.851733+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826401851_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: allocating
ggml_metal_init: found ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC","depth":4,"on_screen":true,"value":"ggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:12:29.052150Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=2 elapsed=2.751484125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-6375099005930742097
|
-852231732216400416
|
visual_change
|
accessibility
|
NULL
|
ggml_metal_init: allocating
ggml_metal_init: found ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:11:22.783747Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:12:29.052150Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=2 elapsed=2.751484125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value ...
|
45865
|
NULL
|
NULL
|
NULL
|
|
45868
|
1640
|
2
|
2026-05-15T06:26:47.907804+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826407907_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","depth":4,"on_screen":true,"value":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
8329854396603190434
|
-851668773673044512
|
visual_change
|
accessibility
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45869
|
1641
|
1
|
2026-05-15T06:27:06.664975+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826426664_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","depth":4,"on_screen":true,"value":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
8329854396603190434
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
45866
|
NULL
|
NULL
|
NULL
|
|
45870
|
1640
|
3
|
2026-05-15T06:27:18.073166+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826438073_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","depth":4,"on_screen":true,"value":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
8329854396603190434
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
45868
|
NULL
|
NULL
|
NULL
|
|
45871
|
NULL
|
0
|
2026-05-15T06:27:36.851860+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826456851_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","depth":4,"on_screen":true,"value":"84125s\n2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
8329854396603190434
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
84125s
2026-05-14T18:12:29.052294Z INFO screenpip 84125s
2026-05-14T18:12:29.052294Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 2 eligible frames
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped ...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45872
|
NULL
|
0
|
2026-05-15T06:27:48.219160+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826468219_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: allocating
ggml_metal_init: found ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) w...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted","depth":4,"on_screen":true,"value":"ggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-3540372755045052228
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: allocating
ggml_metal_init: found ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) w...
|
45868
|
NULL
|
NULL
|
NULL
|
|
45873
|
1643
|
0
|
2026-05-15T06:28:07.019019+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826487019_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy →...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted","depth":4,"on_screen":true,"value":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
6484112855111655431
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy →...
|
45871
|
NULL
|
NULL
|
NULL
|
|
45874
|
1642
|
0
|
2026-05-15T06:28:18.347099+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826498347_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy →...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted","depth":4,"on_screen":true,"value":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
6484112855111655431
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy →...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45875
|
1643
|
1
|
2026-05-15T06:28:37.253310+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826517253_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy →...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted","depth":4,"on_screen":true,"value":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)\n2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2\n2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)\n2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)\n2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)\n2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart\n2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)\n2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-14T18:17:07.444669Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=374 elapsed=3.00804925s\n2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)\n2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)\n2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)\n2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
6484112855111655431
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:13:25.193120Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:15:27.572444Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:04.060905Z INFO screenpipe_engine::sleep_monitor: Screen unlocked (CGSession safety-net poll)
2026-05-14T18:17:04.086797Z INFO screenpipe_engine::event_driven_capture: invalidating persistent streams after unlock/wake for monitor 2
2026-05-14T18:17:04.289459Z INFO sck_rs::stream_manager: stopped 2 persistent stream(s)
2026-05-14T18:17:04.514888Z INFO screenpipe_audio::core::run_record_and_transcribe: screen unlocked — rebuilding stream for MacBook Pro Microphone (input) (avoids zombie-callback state observed after sleep/wake on macOS, Windows, and Linux)
2026-05-14T18:17:04.557979Z WARN screenpipe_audio::audio_manager::manager: recording for device MacBook Pro Microphone (input) exited with error: stream rebuild required after screen unlock for MacBook Pro Microphone (input) (recovery is automatic via device_monitor)
2026-05-14T18:17:05.242244Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-14T18:17:05.549376Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for MacBook Pro Microphone (input), cleaning up for restart
2026-05-14T18:17:05.549451Z INFO screenpipe_audio::device::device_manager: Stopping device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.308691Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-14T18:17:06.846677Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846824Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: MacBook Pro Microphone (input)
2026-05-14T18:17:06.846890Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-14T18:17:07.444669Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=374 elapsed=3.00804925s
2026-05-14T18:17:11.045310Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 160.8ms elapsed (expected 5.3ms) → inserting 155.4ms silence (14923 samples)
2026-05-14T18:17:12.775454Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 83.0ms elapsed (expected 5.3ms) → inserting 77.7ms silence (7460 samples)
2026-05-14T18:17:19.588065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:19.712429Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=visual_change)
2026-05-14T18:17:25.869045Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-4915557953454808779, trigger=click)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy →...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45876
|
1642
|
1
|
2026-05-15T06:28:48.488753+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826528488_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
45874
|
NULL
|
NULL
|
NULL
|
|
45877
|
1643
|
2
|
2026-05-15T06:29:07.474202+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826547474_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
45875
|
NULL
|
NULL
|
NULL
|
|
45878
|
1642
|
2
|
2026-05-15T06:29:18.714557+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826558714_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45879
|
1643
|
3
|
2026-05-15T06:29:37.694610+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826577694_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45880
|
1642
|
3
|
2026-05-15T06:29:48.834522+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826588834_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
45878
|
NULL
|
NULL
|
NULL
|
|
45881
|
1643
|
4
|
2026-05-15T06:30:07.907336+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826607907_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
45879
|
NULL
|
NULL
|
NULL
|
|
45882
|
1642
|
4
|
2026-05-15T06:30:18.964283+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826618964_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
45878
|
NULL
|
NULL
|
NULL
|
|
45883
|
1643
|
5
|
2026-05-15T06:30:38.102252+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826638102_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)\n2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:17:41.655714Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=0 elapsed=12.584709333s\n2026-05-14T18:17:41.659296Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=375 elapsed=1.525763459s\n2026-05-14T18:19:15.613896Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=383 elapsed=1.089600916s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)\n2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-14T18:20:28.676631Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8971157115277618803, trigger=click)\n2026-05-14T18:20:55.977460Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-1564715759782373567, trigger=click)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:21:41.262468Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:22:46.614239Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=0 elapsed=4.94553125s\n2026-05-14T18:22:52.962854Z WARN sqlx::query: summary=\"SELECT DISTINCT app_name, window_name, …\" db.statement=\"\\n\\nSELECT\\n DISTINCT app_name,\\n window_name,\\n browser_url\\nFROM\\n frames\\nWHERE\\n timestamp > datetime('now', '-30 seconds')\\n AND app_name IS NOT NULL\\n AND window_name IS NOT NULL\\n\" rows_affected=0 rows_returned=392 elapsed=1.0767815s\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-7279085127634146319
|
-852231712889047584
|
idle
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:17:32.516386Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 86.0ms elapsed (expected 5.3ms) → inserting 80.7ms silence (7745 samples)
2026-05-14T18:17:32.920163Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:17:41.655714Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=0 elapsed=12.584709333s
2026-05-14T18:17:41.659296Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=375 elapsed=1.525763459s
2026-05-14T18:19:15.613896Z WARN sqlx::query: summary="SELECT DISTINCT app_name, window_name, …" db.statement="\n\nSELECT\n DISTINCT app_name,\n window_name,\n browser_url\nFROM\n frames\nWHERE\n timestamp > datetime('now', '-30 seconds')\n AND app_name IS NOT NULL\n AND window_name IS NOT NULL\n" rows_affected=0 rows_returned=383 elapsed=1.089600916s
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:19:38.508186Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:19:45.114427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
2026-05-14T18:19:55.429434Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=visual_change)
2026-05-14T18:19:57.285192Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-554583693258477392, trigger=click)
tip: install a starter bundle of pipes:
screenpipe install [URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_meta...
|
45879
|
NULL
|
NULL
|
NULL
|
|
45885
|
1643
|
6
|
2026-05-15T06:31:08.303935+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826668303_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_backend_init_gpu: device 0: Metal (type: 1 whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backen...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
5930999014413410901
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
whisper_backend_init_gpu: device 0: Metal (type: 1 whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backen...
|
45879
|
NULL
|
NULL
|
NULL
|
|
45886
|
1642
|
6
|
2026-05-15T06:31:19.224731+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826679224_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_backend_init_gpu: device 0: Metal (type: 1 whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backen...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"whisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
5930999014413410901
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
whisper_backend_init_gpu: device 0: Metal (type: 1 whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backen...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45887
|
1643
|
7
|
2026-05-15T06:31:38.510149+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826698510_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-1477840697668178091
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45888
|
1642
|
7
|
2026-05-15T06:31:49.382409+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826709382_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-1477840697668178091
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
45886
|
NULL
|
NULL
|
NULL
|
|
45889
|
1643
|
8
|
2026-05-15T06:32:08.717711+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826728717_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-1477840697668178091
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
45887
|
NULL
|
NULL
|
NULL
|
|
45890
|
1642
|
8
|
2026-05-15T06:32:19.523639+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826739523_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-1477840697668178091
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45891
|
NULL
|
0
|
2026-05-15T06:32:38.914777+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826758914_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-1477840697668178091
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
45887
|
NULL
|
NULL
|
NULL
|
|
45892
|
NULL
|
0
|
2026-05-15T06:32:49.670909+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826769670_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:27:52.060944Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=17 elapsed=5.421659208s\n2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames\n2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted\n2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nzsh: terminated npx screenpipe@latest record\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start\nsp-start='npx screenpipe@latest record --disable-audio --ignored-windows \"Boosteroid\"'\nlukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record\ndetected hardware tier: Mid\nwarning: parakeet is not supported on this platform, using whisper-tiny instead\n2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store\nchecking permissions...\n screen recording: ok\n microphone: ok\n accessibility: ok\n2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6\n2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor\n2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)\n2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)\n2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true\n2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode\n2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on 127.0.0.1:3030 (localhost only)\n2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key\n2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)\n2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)\n2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager\n2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap\n2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update\n2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits\n2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown\n2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export\n2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary\n2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from \"/Users/lukas/.screenpipe/pipes\"\n\n\n\n _ \n __________________ ___ ____ ____ (_____ ___ \n / ___/ ___/ ___/ _ \\/ _ \\/ __ \\ / __ \\/ / __ \\/ _ \\\n (__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/\n/____/\\___/_/ \\___/\\___/_/ /_/ / .___/_/ .___/\\___/ \n /_/ /_/ \n\n\n\npower AI by everything you've seen, said or heard\n2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)\nopen source | runs locally | developer friendly\n\n\n┌────────────────────────┬────────────────────────────────────┐\n│ setting │ value │\n├────────────────────────┼────────────────────────────────────┤\n│ audio chunk duration │ 30 seconds │\n│ port │ 3030 │\n│ audio disabled │ false │\n│ vision disabled │ false │\n│ pause on DRM content │ false │\n│ audio engine │ Parakeet │\n│ vad engine │ Silero │\n│ data directory │ /Users/lukas/.screenpipe │\n│ debug mode │ false │\n│ telemetry │ true │\n│ use pii removal │ true │\n│ use all monitors │ true │\n│ ignored windows │ [] │\n│ included windows │ [] │\n│ cloud sync │ disabled │\n│ auto-destruct pid │ 0 │\n│ deepgram key │ not set │\n│ api auth │ enabled │\n│ encrypt secrets │ disabled │\n│ retention days │ 14 │\n│ retention mode │ media-only (keep transcripts) │\n├────────────────────────┼────────────────────────────────────┤\n│ languages │ │\n│ │ all languages │\n├────────────────────────┼────────────────────────────────────┤\n│ monitors │ │\n│ │ id: 1 │\n│ │ id: 2 │\n├────────────────────────┼────────────────────────────────────┤\n│ audio devices │ │\n│ │ soundcore AeroClip (input) │\n│ │ System Audio (output) │\n└────────────────────────┴────────────────────────────────────┘\nyou are using local processing. all your data stays on your computer.\n\nwarning: telemetry is enabled. only error-level data will be sent.\nto disable, use the --disable-telemetry flag.\n\ncheck latest changes here: https://github.com/screenpipe/screenpipe/releases\n2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture\n2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh\n2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))\n2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture\n2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.\n2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e\n2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)\n2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)\n2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)\n2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on 127.0.0.1:3030\n2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030\n2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)\n2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)\n2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)\n2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)\n2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)\n2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)\n2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)\n2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)\n2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)\n2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms\n2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms\n2026-05-15T09:26:40.784643Z WARN sqlx::query: summary=\"SELECT f.id, f.timestamp, f.offset_index, …\" db.statement=\"\\n\\nSELECT\\n f.id,\\n f.timestamp,\\n f.offset_index,\\n COALESCE(\\n SUBSTR(f.full_text, 1, 200),\\n SUBSTR(f.accessibility_text, 1, 200),\\n (\\n SELECT\\n SUBSTR(ot.text, 1, 200)\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as text,\\n COALESCE(\\n f.app_name,\\n (\\n SELECT\\n ot.app_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as app_name,\\n COALESCE(\\n f.window_name,\\n (\\n SELECT\\n ot.window_name\\n FROM\\n ocr_text ot\\n WHERE\\n ot.frame_id = f.id\\n LIMIT\\n 1\\n )\\n ) as window_name,\\n COALESCE(vc.device_name, f.device_name) as screen_device,\\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\\n COALESCE(vc.fps, 0.033) as chunk_fps,\\n f.browser_url,\\n f.machine_id\\nFROM\\n frames f\\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\\nWHERE\\n f.timestamp >= ?1\\n AND f.timestamp <= ?2\\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\\nORDER BY\\n f.timestamp DESC,\\n f.offset_index DESC\\nLIMIT\\n 10000\\n\" rows_affected=0 rows_returned=7176 elapsed=5.118563458s\n2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC\n2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: \"/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin\"\n2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)\n2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...\nwhisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'\nwhisper_init_with_params_no_state: use gpu = 1\nwhisper_init_with_params_no_state: flash attn = 0\nwhisper_init_with_params_no_state: gpu_device = 0\nwhisper_init_with_params_no_state: dtw = 0\nggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices\nggml_metal_library_init: using embedded metal library\nggml_metal_library_init: loaded in 0.043 sec\nggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)\nggml_metal_device_init: GPU name: Apple M1\nggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)\nggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)\nggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)\nggml_metal_device_init: simdgroup reduction = true\nggml_metal_device_init: simdgroup matrix mul. = true\nggml_metal_device_init: has unified memory = true\nggml_metal_device_init: has bfloat = true\nggml_metal_device_init: has tensor = false\nggml_metal_device_init: use residency sets = true\nggml_metal_device_init: use shared buffers = true\nggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB\nwhisper_init_with_params_no_state: devices = 3\nwhisper_init_with_params_no_state: backends = 3\nwhisper_model_load: loading model\nwhisper_model_load: n_vocab = 51865\nwhisper_model_load: n_audio_ctx = 1500\nwhisper_model_load: n_audio_state = 384\nwhisper_model_load: n_audio_head = 6\nwhisper_model_load: n_audio_layer = 4\nwhisper_model_load: n_text_ctx = 448\nwhisper_model_load: n_text_state = 384\nwhisper_model_load: n_text_head = 6\nwhisper_model_load: n_text_layer = 4\nwhisper_model_load: n_mels = 80\nwhisper_model_load: ftype = 1\nwhisper_model_load: qntvr = 0\nwhisper_model_load: type = 1 (tiny)\nwhisper_model_load: adding 1608 extra tokens\nwhisper_model_load: n_langs = 99\nwhisper_model_load: Metal total size = 77.11 MB\nwhisper_model_load: model size = 77.11 MB\n2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)\n2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager\n2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started\n2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events\n2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)\n2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)\n2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n2026-05-15T09:27:39.547931Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=209 elapsed=3.914452291s\n2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames\n2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted\n2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted\n2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
-1477840697668178091
|
-852231715036531232
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:23:45.030577Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:24:19.699073Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 81.3ms elapsed (expected 5.3ms) → inserting 75.9ms silence (7288 samples)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:25:49.236566Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:27:52.060944Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=17 elapsed=5.421659208s
2026-05-14T18:27:52.061111Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 17 eligible frames
2026-05-14T18:27:52.632428Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-14T18:27:53.132060Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 9 frames, 1.2MB → 0.6MB (2.2x), 9 JPEGs deleted
2026-05-14T18:27:54.278904Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 8 frames, 1.5MB → 1.1MB (1.4x), 8 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-14T18:29:55.226276Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] record
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ alias sp-start
sp-start='npx screenpipe@latest record --disable-audio --ignored-windows "Boosteroid"'
lukas@Lukas-Kovaliks-MacBook-Pro-Jiminny ~ $ npx screenpipe@latest record
detected hardware tier: Mid
warning: parakeet is not supported on this platform, using whisper-tiny instead
2026-05-15T09:26:34.532591Z INFO screenpipe_engine::auth_key: api auth: key resolved via secret store
checking permissions...
screen recording: ok
microphone: ok
accessibility: ok
2026-05-15T09:26:34.611070Z INFO screenpipe_screen::monitor::macos_version: Detected macOS version: 14.6
2026-05-15T09:26:35.239867Z INFO screenpipe_engine::sleep_monitor: Starting macOS sleep/wake monitor
2026-05-15T09:26:35.241531Z INFO screenpipe_engine::sleep_monitor: Screen lock/unlock observers registered (CFNotificationCenter)
2026-05-15T09:26:35.241930Z INFO screenpipe_engine::sleep_monitor: Display reconfiguration watcher registered (CGDisplayRegisterReconfigurationCallback)
2026-05-15T09:26:35.260543Z INFO screenpipe_engine::permission_monitor: permission monitor started screen=true mic=true accessibility=true keychain=true
2026-05-15T09:26:35.260623Z INFO screenpipe: meeting detector enabled — independent of transcription mode
2026-05-15T09:26:35.632208Z INFO screenpipe: API server listening on [IP_ADDRESS]:3030 (localhost only)
2026-05-15T09:26:35.632265Z INFO screenpipe: API auth enabled — run `screenpipe auth token` to view your key
2026-05-15T09:26:35.632219Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction worker started (min_age=600s, poll=300s)
2026-05-15T09:26:35.632183Z INFO screenpipe_engine::power::manager: power manager started (poll interval: 10s)
2026-05-15T09:26:35.632428Z INFO screenpipe_engine::vision_manager::manager: Starting VisionManager
2026-05-15T09:26:35.634456Z INFO screenpipe_core::pipes: loaded pipe: day-recap
2026-05-15T09:26:35.634681Z INFO screenpipe_core::pipes: loaded pipe: standup-update
2026-05-15T09:26:35.634907Z INFO screenpipe_core::pipes: loaded pipe: ai-habits
2026-05-15T09:26:35.635461Z INFO screenpipe_core::pipes: loaded pipe: time-breakdown
2026-05-15T09:26:35.635547Z INFO screenpipe_core::pipes: loaded pipe: video-export
2026-05-15T09:26:35.635628Z INFO screenpipe_core::pipes: loaded pipe: meeting-summary
2026-05-15T09:26:35.635643Z INFO screenpipe_core::pipes: loaded 6 pipes from "/Users/lukas/.screenpipe/pipes"
_
__________________ ___ ____ ____ (_____ ___
/ ___/ ___/ ___/ _ \/ _ \/ __ \ / __ \/ / __ \/ _ \
(__ / /__/ / / __/ __/ / / / / /_/ / / /_/ / __/
/____/\___/_/ \___/\___/_/ /_/ / .___/_/ .___/\___/
/_/ /_/
power AI by everything you've seen, said or heard
2026-05-15T09:26:35.636459Z INFO screenpipe_core::pipes: pipe scheduler started (generation 2)
open source | runs locally | developer friendly
┌────────────────────────┬────────────────────────────────────┐
│ setting │ value │
├────────────────────────┼────────────────────────────────────┤
│ audio chunk duration │ 30 seconds │
│ port │ 3030 │
│ audio disabled │ false │
│ vision disabled │ false │
│ pause on DRM content │ false │
│ audio engine │ Parakeet │
│ vad engine │ Silero │
│ data directory │ /Users/lukas/.screenpipe │
│ debug mode │ false │
│ telemetry │ true │
│ use pii removal │ true │
│ use all monitors │ true │
│ ignored windows │ [] │
│ included windows │ [] │
│ cloud sync │ disabled │
│ auto-destruct pid │ 0 │
│ deepgram key │ not set │
│ api auth │ enabled │
│ encrypt secrets │ disabled │
│ retention days │ 14 │
│ retention mode │ media-only (keep transcripts) │
├────────────────────────┼────────────────────────────────────┤
│ languages │ │
│ │ all languages │
├────────────────────────┼────────────────────────────────────┤
│ monitors │ │
│ │ id: 1 │
│ │ id: 2 │
├────────────────────────┼────────────────────────────────────┤
│ audio devices │ │
│ │ soundcore AeroClip (input) │
│ │ System Audio (output) │
└────────────────────────┴────────────────────────────────────┘
you are using local processing. all your data stays on your computer.
warning: telemetry is enabled. only error-level data will be sent.
to disable, use the --disable-telemetry flag.
check latest changes here: https://github.com/screenpipe/screenpipe/releases
2026-05-15T09:26:35.637044Z INFO screenpipe: starting UI event capture
2026-05-15T09:26:35.641206Z WARN screenpipe: pi agent install failed: bun not found — install from https://bun.sh
2026-05-15T09:26:35.642634Z INFO screenpipe_engine::power::manager: initial power profile: Performance (on_ac=true, battery=Some(100))
2026-05-15T09:26:35.650899Z INFO screenpipe_engine::ui_recorder: Starting UI event capture
2026-05-15T09:26:35.664987Z INFO screenpipe: text-PII worker skipped at startup — async_pii_redaction=false. OPF model (~2.8 GB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665024Z INFO screenpipe: image-PII worker skipped at startup — async_image_pii_redaction=false. rfdetr_v9 model (~108 MB) will NOT be downloaded or loaded. Toggle via Settings → Privacy → AI PII removal.
2026-05-15T09:26:35.665062Z INFO screenpipe_engine::ui_recorder: UI recording session started: af5dd582-4a7c-40c1-a1c7-bc5cff52b88e
2026-05-15T09:26:35.665217Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warming from DB (2026-05-14 06:26:35.665214 UTC to 2026-05-15 06:26:35.665214 UTC)
2026-05-15T09:26:35.665146Z INFO screenpipe_engine::calendar_speaker_id: speaker identification: started (user_name=<not set>)
2026-05-15T09:26:35.666334Z INFO screenpipe_engine::meeting_detector: meeting v2: detection loop started (base_interval=5s, profiles=12)
2026-05-15T09:26:35.675569Z INFO screenpipe_engine::server: Server listening on [IP_ADDRESS]:3030
2026-05-15T09:26:35.682913Z INFO screenpipe_connect::mdns: mdns: advertising screenpipe on port 3030
2026-05-15T09:26:35.796270Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 1 (1440x900)
2026-05-15T09:26:35.796323Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.796359Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 1 (device: monitor_1)
2026-05-15T09:26:35.852473Z INFO screenpipe_engine::vision_manager::manager: Starting vision recording for monitor 2 (3008x1253)
2026-05-15T09:26:35.852505Z INFO screenpipe_engine::vision_manager::manager: Starting event-driven capture for monitor 2 (device: monitor_2)
2026-05-15T09:26:35.852515Z INFO screenpipe_engine::vision_manager::manager: VisionManager started with 2/2 monitor(s)
2026-05-15T09:26:35.852526Z INFO screenpipe_engine::vision_manager::monitor_watcher: Starting monitor watcher (event-driven via CGDisplayRegisterReconfigurationCallback, 60s backstop poll)
2026-05-15T09:26:35.852543Z INFO screenpipe_engine::event_driven_capture: event-driven capture started for monitor 2 (device: monitor_2)
2026-05-15T09:26:36.408054Z INFO sck_rs::stream_manager: persistent SCK stream started for display 1 (1440x900, 2fps, 0 excluded)
2026-05-15T09:26:36.488401Z INFO sck_rs::stream_manager: persistent SCK stream started for display 2 (3008x1253, 2fps, 0 excluded)
2026-05-15T09:26:36.651362Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 1: frame_id=45865, dur=178ms
2026-05-15T09:26:36.658923Z INFO screenpipe_engine::event_driven_capture: startup capture for monitor 2: frame_id=45866, dur=123ms
2026-05-15T09:26:40.784643Z WARN sqlx::query: summary="SELECT f.id, f.timestamp, f.offset_index, …" db.statement="\n\nSELECT\n f.id,\n f.timestamp,\n f.offset_index,\n COALESCE(\n SUBSTR(f.full_text, 1, 200),\n SUBSTR(f.accessibility_text, 1, 200),\n (\n SELECT\n SUBSTR(ot.text, 1, 200)\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as text,\n COALESCE(\n f.app_name,\n (\n SELECT\n ot.app_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as app_name,\n COALESCE(\n f.window_name,\n (\n SELECT\n ot.window_name\n FROM\n ocr_text ot\n WHERE\n ot.frame_id = f.id\n LIMIT\n 1\n )\n ) as window_name,\n COALESCE(vc.device_name, f.device_name) as screen_device,\n COALESCE(vc.file_path, f.snapshot_path) as video_path,\n COALESCE(vc.fps, 0.033) as chunk_fps,\n f.browser_url,\n f.machine_id\nFROM\n frames f\n LEFT JOIN video_chunks vc ON f.video_chunk_id = vc.id\nWHERE\n f.timestamp >= ?1\n AND f.timestamp <= ?2\n AND COALESCE(vc.file_path, f.snapshot_path, '') NOT LIKE 'cloud://%'\nORDER BY\n f.timestamp DESC,\n f.offset_index DESC\nLIMIT\n 10000\n" rows_affected=0 rows_returned=7176 elapsed=5.118563458s
2026-05-15T09:26:40.807774Z INFO screenpipe_engine::hot_frame_cache: hot_frame_cache: warmed with 7176 frame entries, coverage from 2026-05-14 06:26:35.665214 UTC
2026-05-15T09:26:45.640808Z INFO screenpipe_audio::transcription::engine: whisper model available: "/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin"
2026-05-15T09:26:45.641039Z INFO screenpipe_audio::transcription::whisper::model: whisper context: gpu acceleration enabled (Metal on macOS, Vulkan on Windows)
2026-05-15T09:26:45.641080Z INFO screenpipe_audio::transcription::engine: loading whisper model with GPU acceleration...
whisper_init_from_file_with_params_no_state: loading model from '/Users/lukas/.cache/huggingface/hub/models--ggerganov--whisper.cpp/snapshots/5359861c739e955e79d9a303bcbc70fb988958b1/ggml-tiny.bin'
whisper_init_with_params_no_state: use gpu = 1
whisper_init_with_params_no_state: flash attn = 0
whisper_init_with_params_no_state: gpu_device = 0
whisper_init_with_params_no_state: dtw = 0
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.043 sec
ggml_metal_rsets_init: creating a residency set collection (keep_alive = 180 s)
ggml_metal_device_init: GPU name: Apple M1
ggml_metal_device_init: GPU family: MTLGPUFamilyApple7 (1007)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 11453.25 MB
whisper_init_with_params_no_state: devices = 3
whisper_init_with_params_no_state: backends = 3
whisper_model_load: loading model
whisper_model_load: n_vocab = 51865
whisper_model_load: n_audio_ctx = 1500
whisper_model_load: n_audio_state = 384
whisper_model_load: n_audio_head = 6
whisper_model_load: n_audio_layer = 4
whisper_model_load: n_text_ctx = 448
whisper_model_load: n_text_state = 384
whisper_model_load: n_text_head = 6
whisper_model_load: n_text_layer = 4
whisper_model_load: n_mels = 80
whisper_model_load: ftype = 1
whisper_model_load: qntvr = 0
whisper_model_load: type = 1 (tiny)
whisper_model_load: adding 1608 extra tokens
whisper_model_load: n_langs = 99
whisper_model_load: Metal total size = 77.11 MB
whisper_model_load: model size = 77.11 MB
2026-05-15T09:26:45.767408Z INFO screenpipe_audio::transcription::engine: whisper model loaded successfully
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T09:26:45.771528Z INFO screenpipe_audio::audio_manager::manager: transcription session created (will be reused across segments)
2026-05-15T09:26:45.772132Z INFO screenpipe_audio::audio_manager::manager: seeded 22 speakers (named + unnamed) from DB into embedding manager
2026-05-15T09:26:45.772166Z INFO screenpipe_audio::audio_manager::manager: audio manager started
2026-05-15T09:26:45.772179Z INFO screenpipe_audio::audio_manager::manager: calendar-assisted speaker diarization: listening for meeting events
2026-05-15T09:26:46.236933Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:26:46.369731Z INFO screenpipe_audio::device::device_manager: starting recording for device: System Audio (output)
2026-05-15T09:26:46.369799Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for System Audio (output) (unknown / 30s segments)
2026-05-15T09:26:46.369824Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
2026-05-15T09:27:39.547931Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=209 elapsed=3.914452291s
2026-05-15T09:27:39.548108Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 209 eligible frames
2026-05-15T09:27:44.581971Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 84 frames, 19.6MB → 1.4MB (14.3x), 84 JPEGs deleted
2026-05-15T09:27:50.383561Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 99 frames, 18.0MB → 5.1MB (3.5x), 99 JPEGs deleted
2026-05-15T09:27:52.031179Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 24 frames, 4.1MB → 1.0MB (4.1x), 24 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute ...
|
45890
|
NULL
|
NULL
|
NULL
|
|
45893
|
1645
|
0
|
2026-05-15T06:33:09.086777+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826789086_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
5134857979507764136
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45887
|
NULL
|
NULL
|
NULL
|
|
45894
|
1644
|
0
|
2026-05-15T06:33:19.822707+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826799822_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
5134857979507764136
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45890
|
NULL
|
NULL
|
NULL
|
|
45895
|
1645
|
1
|
2026-05-15T06:33:39.278134+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826819278_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
5134857979507764136
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45896
|
1644
|
1
|
2026-05-15T06:33:49.975042+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826829975_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
5134857979507764136
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45897
|
1645
|
2
|
2026-05-15T06:34:09.485257+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826849485_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
5134857979507764136
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45895
|
NULL
|
NULL
|
NULL
|
|
45898
|
1644
|
2
|
2026-05-15T06:34:20.101889+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826860101_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
5134857979507764136
|
-851668773673044512
|
idle
|
accessibility
|
NULL
|
ggml_metal_free: deallocating
whisper_backend_init ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45896
|
NULL
|
NULL
|
NULL
|
|
45899
|
1644
|
3
|
2026-05-15T06:34:33.709048+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826873709_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","depth":4,"on_screen":true,"value":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
2115983340063954676
|
-852231732216400416
|
visual_change
|
accessibility
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45896
|
NULL
|
NULL
|
NULL
|
|
45900
|
1645
|
3
|
2026-05-15T06:34:39.672461+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826879672_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","depth":4,"on_screen":true,"value":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
2115983340063954676
|
-852231732216400416
|
idle
|
accessibility
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:28:47.974048Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:30:50.199977Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: get the screenpipe desktop app for the full experience
[URL_WITH_CREDENTIALS] (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45895
|
NULL
|
NULL
|
NULL
|
|
45902
|
1645
|
4
|
2026-05-15T06:35:09.860466+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826909860_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
1420047954182519985
|
-815639976654080544
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45903
|
1644
|
5
|
2026-05-15T06:35:25.050580+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826925050_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
1420047954182519985
|
-815639976654080544
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45901
|
NULL
|
NULL
|
NULL
|
|
45904
|
1645
|
5
|
2026-05-15T06:35:40.028126+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826940028_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
1420047954182519985
|
-815639976654080544
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45902
|
NULL
|
NULL
|
NULL
|
|
45905
|
1644
|
6
|
2026-05-15T06:35:55.139695+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826955139_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
1420047954182519985
|
-815639976654080544
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
45906
|
1645
|
6
|
2026-05-15T06:36:10.218862+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826970218_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
1420047954182519985
|
-815639976654080544
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45902
|
NULL
|
NULL
|
NULL
|
|
45907
|
NULL
|
0
|
2026-05-15T06:36:25.279120+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778826985279_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","depth":4,"on_screen":true,"value":"ggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
1420047954182519985
|
-815639976654080544
|
idle
|
accessibility
|
NULL
|
ggml_metal_init: found device: Apple M1
ggml_metal ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45905
|
NULL
|
NULL
|
NULL
|
|
45908
|
NULL
|
0
|
2026-05-15T06:36:40.444662+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778827000444_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity","depth":4,"on_screen":true,"value":"whisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames\n2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect\n2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)\n2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart\n2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)\n2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
-6488007505258613009
|
-851668782262979104
|
idle
|
accessibility
|
NULL
|
whisper_backend_init: using BLAS backend
whisper_i whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:32:52.098771Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 3 eligible frames
2026-05-15T09:32:52.385067Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:34:31.090962Z WARN screenpipe_audio::core::run_record_and_transcribe: no usable audio from soundcore AeroClip (input) for 30s — only zero-fill buffers (likely OS device hijack by another app), triggering reconnect
2026-05-15T09:34:31.093981Z WARN screenpipe_audio::audio_manager::manager: recording for device soundcore AeroClip (input) exited with error: Audio stream zero-fill — no usable data from soundcore AeroClip (input) for 30s (suspected device hijack by another process)
2026-05-15T09:34:32.062369Z WARN screenpipe_audio::audio_manager::device_monitor: [DEVICE_RECOVERY] detected stale recording handle for soundcore AeroClip (input), cleaning up for restart
2026-05-15T09:34:32.062583Z INFO screenpipe_audio::device::device_manager: Stopping device: soundcore AeroClip (input)
2026-05-15T09:34:32.177026Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T09:34:32.177079Z INFO screenpipe_audio::audio_manager::device_monitor: restarted with system default device: soundcore AeroClip (input)
2026-05-15T09:34:32.177108Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:34:54.677866Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
45902
|
NULL
|
NULL
|
NULL
|
|
46072
|
1648
|
28
|
2026-05-15T06:46:39.598038+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778827599598_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.
2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)
2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect
2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)
2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)
2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.\n2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)\n2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect\n2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)\n2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)\n2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.\n2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)\n2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect\n2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)\n2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)\n2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
6994266138650324829
|
-815639976654080544
|
app_switch
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.
2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)
2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect
2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)
2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)
2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
46074
|
1649
|
44
|
2026-05-15T06:46:40.524281+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778827600524_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.
2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)
2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect
2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)
2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)
2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.\n2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)\n2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect\n2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)\n2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)\n2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login","depth":4,"on_screen":true,"value":"whisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)\n2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.\n2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)\n2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)\n2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)\n2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect\n2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)\n2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)\n2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.27027926,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.27227393,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.32928857,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.33128324,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.3882979,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.39029256,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.44730717,"top":1.0,"width":0.05900931,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.44930187,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.5063165,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.50831115,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.56515956,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.5671542,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.62400264,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.62599736,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.6828458,"top":1.0,"width":0.058843084,"height":-0.042298436},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.68484044,"top":1.0,"width":0.005319149,"height":-0.04549086},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.7280585,"top":1.0,"width":0.01861702,"height":-0.023144484},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.49601063,"top":1.0,"width":0.027925532,"height":-0.02394259},"on_screen":true,"role_description":"text"}]...
|
6994266138650324829
|
-815639976654080544
|
app_switch
|
accessibility
|
NULL
|
whisper_init_state: kv self size = 3.15 MB
whi whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:43:04.904386Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:45:06.492820Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T09:45:35.456345Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:35.521370Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-8807740823408271838, trigger=click)
2026-05-15T09:45:44.330889Z WARN screenpipe_audio::core::stream: audio device LakyLak bose qc35 II disconnected. stopping recording.
2026-05-15T09:45:45.308811Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490340Z INFO screenpipe_audio::device::device_manager: starting recording for device: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490468Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: MacBook Pro Microphone (input)
2026-05-15T09:45:45.490644Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for MacBook Pro Microphone (input) (wired / 30s segments)
2026-05-15T09:45:46.568692Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 216.8ms elapsed (expected 5.3ms) → inserting 211.5ms silence (20301 samples)
2026-05-15T09:45:51.835427Z WARN screenpipe_audio::core::run_record_and_transcribe: no audio received from LakyLak bose qc35 II (input) for 8s - stream dead, triggering reconnect
2026-05-15T09:45:53.214324Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=2592904943322971128, trigger=click)
2026-05-15T09:45:59.186781Z INFO screenpipe_engine::meeting_detector: meeting v2: Idle -> Confirming (app=Firefox, signals=2)
2026-05-15T09:46:17.127041Z INFO screenpipe_engine::meeting_detector: meeting v2: Confirming -> Idle (timeout, app=Firefox)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
DOCKER
Close Tab
DEV (-zsh)
Close Tab
APP (-zsh)
Close Tab
-zsh
Close Tab
screenpipe"
Close Tab
ec2-user@ip-10-20-31-146:~ (-zsh)
Close Tab
ec2-user@ip-10-30-129-190:~ (-zsh)
Close Tab
ec2-user@ip-10-30-145-148:~ (-zsh)
Close Tab
⌥⌘1
screenpipe"...
|
NULL
|
NULL
|
NULL
|
NULL
|
|
46384
|
1657
|
0
|
2026-05-15T07:04:17.199274+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778828657199_m2.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_2
|
NULL
|
NULL
|
NULL
|
NULL
|
ActivityLaterMoreJiminny ... ~# jiminny-bg# platfo ActivityLaterMoreJiminny ... ~# jiminny-bg# platform-ticketsic productlaunches# random# releases# sofia-office# support# thank-yous# the_people_of jimi...• Direct messages8. Stoyan TomovP. Galya DimitrovaP. Aneliya Angelova€. Vasil Vasilev8. Stefka Stoyanova DS: Todor Stamatov8. Mario Georgiev.. Nikolay Ivanov2o James Graham "Ro Stoyan Tanev MR. Steliyan Georgiev =f Petko Kashinski. Lukas Kovalik y...i: Apps® Toast$i Jira CloudQ Describe what you are looking forw releases8 22• Messagesr Filese Bookmarks17599/6e18227kd2h66030eebb36d708a924b12dd)-([URL_WITH_CREDENTIALS] accept_deployvcheckout-code 887480build-frontend 887481test-frontend 887482v build-backend 887483phpstan 887484orevare denlov revision_orod 88/4998build_docker_backend_code_prod 887502build docker worker code prod 8875008 build_docker_worker_video_code_prod 88750db_migrations_prod 8875038 deploy_frontend_assets_to_s3_prod 8875048denlov docker backend code_ orod 887506denlov docker worker code nrod 887505deploy_docker_worker_video_code_prod 8875078sentrv notiiv-1 887508nronaro denlov rovicion nrod eu 8871878build_docker_backend_code_prod_eu 887489All statusesCheckout courcemcp-tools-schemas30252/b Merge branch 'master into mcp-tools-schemasmasteraOOd53b Merge pull request #12073 from jiminny/JY-20891-improve-sms-text-relays E40 ll"Jiminny lech Day • now100% L28• Fri 15 May 10:04:16Trigger eventPush Commit pushedPush Commit pushedStart1m ago4m ago4 Manage triggersnuration59s57s3m 33s1m 12s1m 30s41S1m 19s43sTrigger Pipeline= Display optionsActionsGG.....
|
NULL
|
5256008994648596099
|
NULL
|
visual_change
|
ocr
|
NULL
|
ActivityLaterMoreJiminny ... ~# jiminny-bg# platfo ActivityLaterMoreJiminny ... ~# jiminny-bg# platform-ticketsic productlaunches# random# releases# sofia-office# support# thank-yous# the_people_of jimi...• Direct messages8. Stoyan TomovP. Galya DimitrovaP. Aneliya Angelova€. Vasil Vasilev8. Stefka Stoyanova DS: Todor Stamatov8. Mario Georgiev.. Nikolay Ivanov2o James Graham "Ro Stoyan Tanev MR. Steliyan Georgiev =f Petko Kashinski. Lukas Kovalik y...i: Apps® Toast$i Jira CloudQ Describe what you are looking forw releases8 22• Messagesr Filese Bookmarks17599/6e18227kd2h66030eebb36d708a924b12dd)-([URL_WITH_CREDENTIALS] accept_deployvcheckout-code 887480build-frontend 887481test-frontend 887482v build-backend 887483phpstan 887484orevare denlov revision_orod 88/4998build_docker_backend_code_prod 887502build docker worker code prod 8875008 build_docker_worker_video_code_prod 88750db_migrations_prod 8875038 deploy_frontend_assets_to_s3_prod 8875048denlov docker backend code_ orod 887506denlov docker worker code nrod 887505deploy_docker_worker_video_code_prod 8875078sentrv notiiv-1 887508nronaro denlov rovicion nrod eu 8871878build_docker_backend_code_prod_eu 887489All statusesCheckout courcemcp-tools-schemas30252/b Merge branch 'master into mcp-tools-schemasmasteraOOd53b Merge pull request #12073 from jiminny/JY-20891-improve-sms-text-relays E40 ll"Jiminny lech Day • now100% L28• Fri 15 May 10:04:16Trigger eventPush Commit pushedPush Commit pushedStart1m ago4m ago4 Manage triggersnuration59s57s3m 33s1m 12s1m 30s41S1m 19s43sTrigger Pipeline= Display optionsActionsGG.....
|
NULL
|
NULL
|
NULL
|
NULL
|
|
46385
|
1656
|
1
|
2026-05-15T07:04:18.934795+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778828658934_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:59:25.294998Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:00:16.744300Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)
2026-05-15T10:00:22.419896Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:00:25.805165Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks
2026-05-15T10:00:35.500672Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8245048993371630268, trigger=visual_change)
2026-05-15T10:00:41.940615Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Active (controls reappeared, app=Firefox, id=9, hysteresis=2/2)
2026-05-15T10:01:14.982726Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:01:28.067481Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-15T10:01:54.206890Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:03.337683Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:06.366420Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:12.424846Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:15.417860Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:33.622277Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)
2026-05-15T10:02:41.504330Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 136.8ms elapsed (expected 4.6ms) → inserting 132.2ms silence (12691 samples)
2026-05-15T10:02:50.510031Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:03:31.131209Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=112 elapsed=2.910282333s
2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames
2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted
2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted
2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)
2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 ...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:59:25.294998Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:00:16.744300Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)\n2026-05-15T10:00:22.419896Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:00:25.805165Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks\n2026-05-15T10:00:35.500672Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8245048993371630268, trigger=visual_change)\n2026-05-15T10:00:41.940615Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Active (controls reappeared, app=Firefox, id=9, hysteresis=2/2)\n2026-05-15T10:01:14.982726Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:01:28.067481Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-15T10:01:54.206890Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:03.337683Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:06.366420Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:12.424846Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:15.417860Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:33.622277Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)\n2026-05-15T10:02:41.504330Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 136.8ms elapsed (expected 4.6ms) → inserting 132.2ms silence (12691 samples)\n2026-05-15T10:02:50.510031Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:03:31.131209Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=112 elapsed=2.910282333s\n2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames\n2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted\n2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted\n2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)\n2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)","depth":4,"on_screen":true,"value":"ggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T09:59:25.294998Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:00:16.744300Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)\n2026-05-15T10:00:22.419896Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:00:25.805165Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks\n2026-05-15T10:00:35.500672Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8245048993371630268, trigger=visual_change)\n2026-05-15T10:00:41.940615Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Active (controls reappeared, app=Firefox, id=9, hysteresis=2/2)\n2026-05-15T10:01:14.982726Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:01:28.067481Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\n2026-05-15T10:01:54.206890Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:03.337683Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:06.366420Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:12.424846Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:15.417860Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:02:33.622277Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)\n2026-05-15T10:02:41.504330Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 136.8ms elapsed (expected 4.6ms) → inserting 132.2ms silence (12691 samples)\n2026-05-15T10:02:50.510031Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:03:31.131209Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=112 elapsed=2.910282333s\n2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames\n2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted\n2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted\n2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)\n2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
3475821339256002975
|
-815639959474211360
|
visual_change
|
accessibility
|
NULL
|
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T09:59:25.294998Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:00:16.744300Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)
2026-05-15T10:00:22.419896Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:00:25.805165Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks
2026-05-15T10:00:35.500672Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8245048993371630268, trigger=visual_change)
2026-05-15T10:00:41.940615Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Active (controls reappeared, app=Firefox, id=9, hysteresis=2/2)
2026-05-15T10:01:14.982726Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:01:28.067481Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
tip: install a starter bundle of pipes:
screenpipe install https://screenpi.pe/start.json
2026-05-15T10:01:54.206890Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:03.337683Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:06.366420Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:12.424846Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:15.417860Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:02:33.622277Z INFO screenpipe_engine::meeting_detector: meeting v2: Active -> Ending (no controls, app=Firefox, id=9, grace=300s)
2026-05-15T10:02:41.504330Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 136.8ms elapsed (expected 4.6ms) → inserting 132.2ms silence (12691 samples)
2026-05-15T10:02:50.510031Z INFO screenpipe_audio::audio_manager::manager: batch mode: audio session ended, transcribing accumulated audio
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:03:31.131209Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=112 elapsed=2.910282333s
2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames
2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted
2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted
2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)
2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 ...
|
46383
|
NULL
|
NULL
|
NULL
|
|
46722
|
1664
|
6
|
2026-05-15T07:24:50.445406+00:00
|
/Users/lukas/.screenpipe/data/data/2026-05-15/1778 /Users/lukas/.screenpipe/data/data/2026-05-15/1778829890445_m1.jpg...
|
iTerm2
|
screenpipe"
|
1
|
NULL
|
monitor_1
|
NULL
|
NULL
|
NULL
|
NULL
|
whisper_init_state: kv cross size = 9.44 MB
whi whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:03:31.131209Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=112 elapsed=2.910282333s
2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames
2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted
2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted
2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)
2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:04:45.749136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6964559813879236637, trigger=click)
2026-05-15T10:05:27.172337Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=visual_change)
2026-05-15T10:05:27.510755Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=click)
2026-05-15T10:05:30.812806Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 88.3ms elapsed (expected 5.3ms) → inserting 83.0ms silence (7967 samples)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:05:31.079370Z WARN screenpipe_audio::core::stream: audio device soundcore AeroClip disconnected. stopping recording.
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:05:31.972705Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:05:33.134587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8471196323029320898, trigger=visual_change)
2026-05-15T10:05:33.295828Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: soundcore AeroClip (input)
2026-05-15T10:05:33.298197Z INFO screenpipe_audio::core::run_record_and_transcribe: stopped recording for MacBook Pro Microphone (input)
2026-05-15T10:05:33.391043Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T10:05:33.391518Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: soundcore AeroClip (input)
2026-05-15T10:05:33.391563Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
2026-05-15T10:06:34.921783Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:36.545159Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:39.578502Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:42.595977Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:48.684301Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:57.733179Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:07:00.736489Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:07:03.787465Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:07:14.033965Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:07:26.111577Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:07:33.533196Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:07:35.698235Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Idle (timeout=300s, app=Firefox, id=9)
2026-05-15T10:07:35.875217Z INFO screenpipe_engine::meeting_detector: meeting v2: meeting ended (id=9)
2026-05-15T10:07:35.875613Z INFO screenpipe_engine::routes::meetings: meeting_status_changed: active=false, manual=false, active_id=None, app=None, source=None
2026-05-15T10:07:50.305447Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:07:56.377587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:02.475230Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:26.625208Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:32.685136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:35.692428Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:40.793299Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=72 elapsed=2.194120875s
2026-05-15T10:08:40.793539Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 72 eligible frames
2026-05-15T10:08:43.904052Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 48 frames, 8.0MB → 2.9MB (2.7x), 48 JPEGs deleted
2026-05-15T10:08:45.460955Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 3.1MB → 0.7MB (4.7x), 22 JPEGs deleted
2026-05-15T10:08:55.698065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:09:35.087072Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:09:47.734757Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:10:13.880551Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)
2026-05-15T10:10:16.919796Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)
2026-05-15T10:10:41.104797Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=179898036855498210, trigger=visual_change)
2026-05-15T10:10:54.258773Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1154276312842020795, trigger=click)
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:11:37.577060Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:11:46.933038Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=6120817221547753544, trigger=visual_change)
2026-05-15T10:11:49.145157Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)
2026-05-15T10:11:58.197330Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:13:39.002363Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:13:49.889227Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=1 rows_returned=84 elapsed=4.427893417s
2026-05-15T10:13:49.889350Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 84 eligible frames
2026-05-15T10:13:52.929679Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 50 frames, 7.8MB → 3.3MB (2.3x), 50 JPEGs deleted
2026-05-15T10:13:55.264589Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 32 frames, 4.9MB → 2.0MB (2.5x), 32 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:15:40.418336Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:16:25.815257Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:16:28.831240Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
2026-05-15T10:17:24.077685Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:17:26.705227Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:17:27.051451Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:17:29.707518Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T10:17:41.889109Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:17:42.150435Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:17:44.896908Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:18:55.686978Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 78 eligible frames
2026-05-15T10:18:58.946619Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 8.0MB → 3.7MB (2.2x), 47 JPEGs deleted
2026-05-15T10:19:01.162533Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 29 frames, 4.6MB → 1.5MB (3.0x), 29 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:19:43.534442Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:19:54.454063Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=293848048195837472, trigger=click)
tip: install a starter bundle of pip...
|
[{"role":"AXTextArea","text [{"role":"AXTextArea","text":"whisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:03:31.131209Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=112 elapsed=2.910282333s\n2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames\n2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted\n2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted\n2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)\n2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:04:45.749136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6964559813879236637, trigger=click)\n2026-05-15T10:05:27.172337Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=visual_change)\n2026-05-15T10:05:27.510755Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=click)\n2026-05-15T10:05:30.812806Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 88.3ms elapsed (expected 5.3ms) → inserting 83.0ms silence (7967 samples)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:05:31.079370Z WARN screenpipe_audio::core::stream: audio device soundcore AeroClip disconnected. stopping recording.\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:05:31.972705Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:05:33.134587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8471196323029320898, trigger=visual_change)\n2026-05-15T10:05:33.295828Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: soundcore AeroClip (input)\n2026-05-15T10:05:33.298197Z INFO screenpipe_audio::core::run_record_and_transcribe: stopped recording for MacBook Pro Microphone (input)\n2026-05-15T10:05:33.391043Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T10:05:33.391518Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: soundcore AeroClip (input)\n2026-05-15T10:05:33.391563Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\n2026-05-15T10:06:34.921783Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:36.545159Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:39.578502Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:42.595977Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:48.684301Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:57.733179Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:07:00.736489Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:07:03.787465Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:07:14.033965Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:07:26.111577Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:07:33.533196Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:07:35.698235Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Idle (timeout=300s, app=Firefox, id=9)\n2026-05-15T10:07:35.875217Z INFO screenpipe_engine::meeting_detector: meeting v2: meeting ended (id=9)\n2026-05-15T10:07:35.875613Z INFO screenpipe_engine::routes::meetings: meeting_status_changed: active=false, manual=false, active_id=None, app=None, source=None\n2026-05-15T10:07:50.305447Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:07:56.377587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:02.475230Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:26.625208Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:32.685136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:35.692428Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:40.793299Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=72 elapsed=2.194120875s\n2026-05-15T10:08:40.793539Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 72 eligible frames\n2026-05-15T10:08:43.904052Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 48 frames, 8.0MB → 2.9MB (2.7x), 48 JPEGs deleted\n2026-05-15T10:08:45.460955Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 3.1MB → 0.7MB (4.7x), 22 JPEGs deleted\n2026-05-15T10:08:55.698065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:09:35.087072Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:09:47.734757Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:10:13.880551Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)\n2026-05-15T10:10:16.919796Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)\n2026-05-15T10:10:41.104797Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=179898036855498210, trigger=visual_change)\n2026-05-15T10:10:54.258773Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1154276312842020795, trigger=click)\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:11:37.577060Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:11:46.933038Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=6120817221547753544, trigger=visual_change)\n2026-05-15T10:11:49.145157Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)\n2026-05-15T10:11:58.197330Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:13:39.002363Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:13:49.889227Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=84 elapsed=4.427893417s\n2026-05-15T10:13:49.889350Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 84 eligible frames\n2026-05-15T10:13:52.929679Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 50 frames, 7.8MB → 3.3MB (2.3x), 50 JPEGs deleted\n2026-05-15T10:13:55.264589Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 32 frames, 4.9MB → 2.0MB (2.5x), 32 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:15:40.418336Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:16:25.815257Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:16:28.831240Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\n2026-05-15T10:17:24.077685Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:17:26.705227Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:17:27.051451Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:17:29.707518Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T10:17:41.889109Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:17:42.150435Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:17:44.896908Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:18:55.686978Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 78 eligible frames\n2026-05-15T10:18:58.946619Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 8.0MB → 3.7MB (2.2x), 47 JPEGs deleted\n2026-05-15T10:19:01.162533Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 29 frames, 4.6MB → 1.5MB (3.0x), 29 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:19:43.534442Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:19:54.454063Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=293848048195837472, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:21:45.045347Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:23:44.091728Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=3783661799407507212, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:23:46.406743Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:23:47.100220Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=3783661799407507212, trigger=visual_change)\n2026-05-15T10:23:50.168615Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=3783661799407507212, trigger=visual_change)\n2026-05-15T10:23:56.209427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-3131380257679146262, trigger=visual_change)\n2026-05-15T10:24:01.598840Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 83 eligible frames\n2026-05-15T10:24:03.468500Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 30 frames, 4.0MB → 0.8MB (5.1x), 30 JPEGs deleted\n2026-05-15T10:24:06.843609Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 51 frames, 9.5MB → 3.2MB (2.9x), 51 JPEGs deleted\n2026-05-15T10:24:14.439978Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7555493398188147089, trigger=visual_change)","depth":4,"on_screen":true,"value":"whisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:03:31.131209Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=112 elapsed=2.910282333s\n2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames\n2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted\n2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted\n2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)\n2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)\n2026-05-15T10:04:45.749136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6964559813879236637, trigger=click)\n2026-05-15T10:05:27.172337Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=visual_change)\n2026-05-15T10:05:27.510755Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=click)\n2026-05-15T10:05:30.812806Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 88.3ms elapsed (expected 5.3ms) → inserting 83.0ms silence (7967 samples)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:05:31.079370Z WARN screenpipe_audio::core::stream: audio device soundcore AeroClip disconnected. stopping recording.\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:05:31.972705Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:05:33.134587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8471196323029320898, trigger=visual_change)\n2026-05-15T10:05:33.295828Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: soundcore AeroClip (input)\n2026-05-15T10:05:33.298197Z INFO screenpipe_audio::core::run_record_and_transcribe: stopped recording for MacBook Pro Microphone (input)\n2026-05-15T10:05:33.391043Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)\n2026-05-15T10:05:33.391518Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: soundcore AeroClip (input)\n2026-05-15T10:05:33.391563Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)\n\n tip: sign in for higher AI quotas + cloud sync:\n screenpipe login\n\n2026-05-15T10:06:34.921783Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:36.545159Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:39.578502Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:42.595977Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:48.684301Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:06:57.733179Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:07:00.736489Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:07:03.787465Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)\n2026-05-15T10:07:14.033965Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:07:26.111577Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:07:33.533196Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:07:35.698235Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Idle (timeout=300s, app=Firefox, id=9)\n2026-05-15T10:07:35.875217Z INFO screenpipe_engine::meeting_detector: meeting v2: meeting ended (id=9)\n2026-05-15T10:07:35.875613Z INFO screenpipe_engine::routes::meetings: meeting_status_changed: active=false, manual=false, active_id=None, app=None, source=None\n2026-05-15T10:07:50.305447Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:07:56.377587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:02.475230Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:26.625208Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:32.685136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:35.692428Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:08:40.793299Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=0 rows_returned=72 elapsed=2.194120875s\n2026-05-15T10:08:40.793539Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 72 eligible frames\n2026-05-15T10:08:43.904052Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 48 frames, 8.0MB → 2.9MB (2.7x), 48 JPEGs deleted\n2026-05-15T10:08:45.460955Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 3.1MB → 0.7MB (4.7x), 22 JPEGs deleted\n2026-05-15T10:08:55.698065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:09:35.087072Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:09:47.734757Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)\n2026-05-15T10:10:13.880551Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)\n2026-05-15T10:10:16.919796Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)\n2026-05-15T10:10:41.104797Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=179898036855498210, trigger=visual_change)\n2026-05-15T10:10:54.258773Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1154276312842020795, trigger=click)\n\n tip: get the screenpipe desktop app for the full experience\n https://screenpi.pe\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:11:37.577060Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:11:46.933038Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=6120817221547753544, trigger=visual_change)\n2026-05-15T10:11:49.145157Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)\n2026-05-15T10:11:58.197330Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:13:39.002363Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:13:49.889227Z WARN sqlx::query: summary=\"SELECT id, snapshot_path, device_name, …\" db.statement=\"\\n\\nSELECT\\n id,\\n snapshot_path,\\n device_name,\\n timestamp\\nFROM\\n frames\\nWHERE\\n snapshot_path IS NOT NULL\\n AND timestamp < ?1\\nORDER BY\\n device_name,\\n timestamp ASC\\nLIMIT\\n 5000\\n\" rows_affected=1 rows_returned=84 elapsed=4.427893417s\n2026-05-15T10:13:49.889350Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 84 eligible frames\n2026-05-15T10:13:52.929679Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 50 frames, 7.8MB → 3.3MB (2.3x), 50 JPEGs deleted\n2026-05-15T10:13:55.264589Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 32 frames, 4.9MB → 2.0MB (2.5x), 32 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:15:40.418336Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:16:25.815257Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:16:28.831240Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n\n tip: wire screenpipe into claude with one command:\n claude mcp add screenpipe -- npx -y screenpipe-mcp\n then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity\n\n2026-05-15T10:17:24.077685Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:17:26.705227Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:17:27.051451Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:17:29.707518Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\n2026-05-15T10:17:41.889109Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:17:42.150435Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:17:44.896908Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)\n2026-05-15T10:18:55.686978Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 78 eligible frames\n2026-05-15T10:18:58.946619Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 8.0MB → 3.7MB (2.2x), 47 JPEGs deleted\n2026-05-15T10:19:01.162533Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 29 frames, 4.6MB → 1.5MB (3.0x), 29 JPEGs deleted\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:19:43.534442Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:19:54.454063Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=293848048195837472, trigger=click)\n\n tip: install a starter bundle of pipes:\n screenpipe install https://screenpi.pe/start.json\n\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:21:45.045347Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:23:44.091728Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=3783661799407507212, trigger=visual_change)\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\nwhisper_backend_init_gpu: device 0: Metal (type: 1)\nwhisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)\nwhisper_backend_init_gpu: using Metal backend\nggml_metal_init: allocating\nggml_metal_init: found device: Apple M1\nggml_metal_init: picking default device: Apple M1\nggml_metal_init: use fusion = true\nggml_metal_init: use concurrency = true\nggml_metal_init: use graph optimize = true\nwhisper_backend_init: using BLAS backend\nwhisper_init_state: kv self size = 3.15 MB\nwhisper_init_state: kv cross size = 9.44 MB\nwhisper_init_state: kv pad size = 2.36 MB\nwhisper_init_state: compute buffer (conv) = 14.17 MB\nwhisper_init_state: compute buffer (encode) = 65.96 MB\nwhisper_init_state: compute buffer (cross) = 8.50 MB\nwhisper_init_state: compute buffer (decode) = 96.83 MB\nggml_metal_free: deallocating\n2026-05-15T10:23:46.406743Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks\n2026-05-15T10:23:47.100220Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=3783661799407507212, trigger=visual_change)\n2026-05-15T10:23:50.168615Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=3783661799407507212, trigger=visual_change)\n2026-05-15T10:23:56.209427Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-3131380257679146262, trigger=visual_change)\n2026-05-15T10:24:01.598840Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 83 eligible frames\n2026-05-15T10:24:03.468500Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 30 frames, 4.0MB → 0.8MB (5.1x), 30 JPEGs deleted\n2026-05-15T10:24:06.843609Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 51 frames, 9.5MB → 3.2MB (2.9x), 51 JPEGs deleted\n2026-05-15T10:24:14.439978Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7555493398188147089, trigger=visual_change)","is_focused":true},{"role":"AXRadioButton","text":"DOCKER","depth":2,"bounds":{"left":0.0,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.004166667,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"DEV (-zsh)","depth":2,"bounds":{"left":0.12326389,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.12743056,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"APP (-zsh)","depth":2,"bounds":{"left":0.24652778,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.25069445,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"-zsh","depth":2,"bounds":{"left":0.36979166,"top":0.05888889,"width":0.12326389,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.37395832,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"screenpipe\"","depth":2,"bounds":{"left":0.49305555,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.49722221,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-20-31-146:~ (-zsh)","depth":2,"bounds":{"left":0.6159722,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.6201389,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-129-190:~ (-zsh)","depth":2,"bounds":{"left":0.73888886,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.7430556,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXRadioButton","text":"ec2-user@ip-10-30-145-148:~ (-zsh)","depth":2,"bounds":{"left":0.86180556,"top":0.05888889,"width":0.12291667,"height":0.026666667},"on_screen":true,"role_description":"radio button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXButton","text":"Close Tab","depth":3,"bounds":{"left":0.8659722,"top":0.06333333,"width":0.011111111,"height":0.017777778},"on_screen":true,"role_description":"button","is_enabled":false,"is_focused":false,"is_selected":false,"is_expanded":false},{"role":"AXStaticText","text":"⌥⌘1","depth":1,"bounds":{"left":0.95625,"top":0.032222223,"width":0.03888889,"height":0.018888889},"on_screen":true,"automation_id":"_NS:8","role_description":"text"},{"role":"AXStaticText","text":"screenpipe\"","depth":1,"bounds":{"left":0.47152779,"top":0.033333335,"width":0.058333334,"height":0.017777778},"on_screen":true,"role_description":"text"}]...
|
8363348460165472767
|
-816194660942940703
|
visual_change
|
accessibility
|
NULL
|
whisper_init_state: kv cross size = 9.44 MB
whi whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:02:51.961754Z INFO screenpipe_audio::audio_manager::manager: batch mode: transcribed 50 chunks
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:03:29.780545Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:03:31.131209Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=112 elapsed=2.910282333s
2026-05-15T10:03:31.131509Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 112 eligible frames
2026-05-15T10:03:35.912818Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 71 frames, 12.8MB → 5.5MB (2.3x), 71 JPEGs deleted
2026-05-15T10:03:38.599618Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 39 frames, 6.5MB → 1.9MB (3.5x), 39 JPEGs deleted
2026-05-15T10:03:58.114269Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=click)
2026-05-15T10:04:02.242059Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:04:16.055543Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7514941666263664050, trigger=visual_change)
2026-05-15T10:04:45.749136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6964559813879236637, trigger=click)
2026-05-15T10:05:27.172337Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=visual_change)
2026-05-15T10:05:27.510755Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-6968430965336078127, trigger=click)
2026-05-15T10:05:30.812806Z WARN screenpipe_audio::core::source_buffer: [MacBook Pro Microphone (input)] large gap on wired device: 88.3ms elapsed (expected 5.3ms) → inserting 83.0ms silence (7967 samples)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:05:31.079370Z WARN screenpipe_audio::core::stream: audio device soundcore AeroClip disconnected. stopping recording.
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:05:31.972705Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:05:33.134587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8471196323029320898, trigger=visual_change)
2026-05-15T10:05:33.295828Z INFO screenpipe_audio::audio_manager::device_monitor: system default input changed to: soundcore AeroClip (input)
2026-05-15T10:05:33.298197Z INFO screenpipe_audio::core::run_record_and_transcribe: stopped recording for MacBook Pro Microphone (input)
2026-05-15T10:05:33.391043Z INFO screenpipe_audio::device::device_manager: starting recording for device: soundcore AeroClip (input)
2026-05-15T10:05:33.391518Z INFO screenpipe_audio::audio_manager::device_monitor: switched to new system default input: soundcore AeroClip (input)
2026-05-15T10:05:33.391563Z INFO screenpipe_audio::core::run_record_and_transcribe: starting continuous recording for soundcore AeroClip (input) (bluetooth / 30s segments)
tip: sign in for higher AI quotas + cloud sync:
screenpipe login
2026-05-15T10:06:34.921783Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:36.545159Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:39.578502Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:42.595977Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:48.684301Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:06:57.733179Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:07:00.736489Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:07:03.787465Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=8592070119316887513, trigger=visual_change)
2026-05-15T10:07:14.033965Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:07:26.111577Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:07:33.533196Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:07:35.698235Z INFO screenpipe_engine::meeting_detector: meeting v2: Ending -> Idle (timeout=300s, app=Firefox, id=9)
2026-05-15T10:07:35.875217Z INFO screenpipe_engine::meeting_detector: meeting v2: meeting ended (id=9)
2026-05-15T10:07:35.875613Z INFO screenpipe_engine::routes::meetings: meeting_status_changed: active=false, manual=false, active_id=None, app=None, source=None
2026-05-15T10:07:50.305447Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:07:56.377587Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:02.475230Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:26.625208Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:32.685136Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:35.692428Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:08:40.793299Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=0 rows_returned=72 elapsed=2.194120875s
2026-05-15T10:08:40.793539Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 72 eligible frames
2026-05-15T10:08:43.904052Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 48 frames, 8.0MB → 2.9MB (2.7x), 48 JPEGs deleted
2026-05-15T10:08:45.460955Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 22 frames, 3.1MB → 0.7MB (4.7x), 22 JPEGs deleted
2026-05-15T10:08:55.698065Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:09:35.087072Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:09:47.734757Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=7118676075647864272, trigger=visual_change)
2026-05-15T10:10:13.880551Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)
2026-05-15T10:10:16.919796Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6670048487670008310, trigger=visual_change)
2026-05-15T10:10:41.104797Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=179898036855498210, trigger=visual_change)
2026-05-15T10:10:54.258773Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=1154276312842020795, trigger=click)
tip: get the screenpipe desktop app for the full experience
https://screenpi.pe
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:11:37.577060Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:11:46.933038Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=6120817221547753544, trigger=visual_change)
2026-05-15T10:11:49.145157Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)
2026-05-15T10:11:58.197330Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=6120817221547753544, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:13:39.002363Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:13:49.889227Z WARN sqlx::query: summary="SELECT id, snapshot_path, device_name, …" db.statement="\n\nSELECT\n id,\n snapshot_path,\n device_name,\n timestamp\nFROM\n frames\nWHERE\n snapshot_path IS NOT NULL\n AND timestamp < ?1\nORDER BY\n device_name,\n timestamp ASC\nLIMIT\n 5000\n" rows_affected=1 rows_returned=84 elapsed=4.427893417s
2026-05-15T10:13:49.889350Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 84 eligible frames
2026-05-15T10:13:52.929679Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 50 frames, 7.8MB → 3.3MB (2.3x), 50 JPEGs deleted
2026-05-15T10:13:55.264589Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 32 frames, 4.9MB → 2.0MB (2.5x), 32 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:15:40.418336Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:16:25.815257Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:16:28.831240Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
tip: wire screenpipe into claude with one command:
claude mcp add screenpipe -- npx -y screenpipe-mcp
then ask claude to build a pipe that tracks who you are, your todos, and how you spend your time from your screen activity
2026-05-15T10:17:24.077685Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:17:26.705227Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:17:27.051451Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:17:29.707518Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
2026-05-15T10:17:41.889109Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:17:42.150435Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:17:44.896908Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 1 (hash=-5229934648328219625, trigger=visual_change)
2026-05-15T10:18:55.686978Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: found 78 eligible frames
2026-05-15T10:18:58.946619Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 47 frames, 8.0MB → 3.7MB (2.2x), 47 JPEGs deleted
2026-05-15T10:19:01.162533Z INFO screenpipe_engine::snapshot_compaction: snapshot compaction: 29 frames, 4.6MB → 1.5MB (3.0x), 29 JPEGs deleted
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
whisper_backend_init_gpu: device 0: Metal (type: 1)
whisper_backend_init_gpu: found GPU device 0: Metal (type: 1, cnt: 0)
whisper_backend_init_gpu: using Metal backend
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M1
ggml_metal_init: picking default device: Apple M1
ggml_metal_init: use fusion = true
ggml_metal_init: use concurrency = true
ggml_metal_init: use graph optimize = true
whisper_backend_init: using BLAS backend
whisper_init_state: kv self size = 3.15 MB
whisper_init_state: kv cross size = 9.44 MB
whisper_init_state: kv pad size = 2.36 MB
whisper_init_state: compute buffer (conv) = 14.17 MB
whisper_init_state: compute buffer (encode) = 65.96 MB
whisper_init_state: compute buffer (cross) = 8.50 MB
whisper_init_state: compute buffer (decode) = 96.83 MB
ggml_metal_free: deallocating
2026-05-15T10:19:43.534442Z INFO screenpipe_audio::audio_manager::manager: reconciliation: transcribed 50 orphaned chunks
2026-05-15T10:19:54.454063Z INFO screenpipe_engine::event_driven_capture: content dedup: skipping capture for monitor 2 (hash=293848048195837472, trigger=click)
tip: install a starter bundle of pip...
|
NULL
|
NULL
|
NULL
|
NULL
|